STUDY ON THE DIMENSIONALITY OF THE COPING INVENTORY(COPE)
Several preliminary studies focused on the adaptation of the coping inventory (COPE) developed by C.S.Carver, et al to assess the coping behavior of Chinese college students. However, there existed the problem of inconsistency in factor structure among the instrument versions in source and target languages,and the revised Chinese version (C-COPE) was consequently reported. The current study was to further examine the dimensionality of COPE and C-COPE. In study 1, second-order exploratory factor analysis was conducted to explore the hierarchical dimensionality of C-COPE on the data collected in the sample of 736 college students, and derived a model which was subsequently tested. In study 2, ten priori models were assumed on the basis of previous researches, among which two models of fourteen first-order factors and four models of four second-order factors were derived from the COPE inventory, two models of eight first-order factors and two models of three second-order factors were derived from the C-COPE inventory. With the LISREL VIII, the confirmatory factor analysis was completed to test the goodness-of-fit of these models to the assessment data from another sample of college students (N＝465). As the multiple indexes (i.e. chi-squire, GFI, AGFI, RMR, NFI, NNFI, CFI, IFI) and the significant difference of chi-squire between nested models indicated, the result supported the oblique first-order-factor model of the C-COPE assessing the eight coping dimensions (i.e. seeking social support, active control and planning, accumulation of mental resources, positive appraisal, focus on and venting of emotions, denial and mental disengagement, superstition coping, helplessness reaction). The hierarchical dimensionality of both COPE and C-COPE was not confirmed because of the instability of second-order factor structures across assessments in different samples. The conclusion drawn from the empirical evidence as well as theoretical analysis in this study was that there did exist construct inequivalence in cross-cultural assessment of coping by the COPE inventory and therefore adaptation approach instead of application of this instrument should be adopted, and that the multidimensional model of C-COPE was more suitable for assessing coping behaviors people adopted in Chinese society. The construct bias as well as item bias was analyzed in detail. The approach of improving the psychometric properties of C-COPE was also discussed.